Note: Most Internet Explorer 8 users encounter issues playing the presentation videos. Please update your browser or use a different one if available.

The Role of Declarative Memory Process for Children with Low Functioning Autism When Using Personalised Speech Generated Devices

Friday, 3 May 2013: 14:00-18:00
Banquet Hall (Kursaal Centre)
S. J. ni Chuileann, Psychology, Trinity College Dublin, Co. Carlow, Ireland
The role of declarative memory process for children with low functioning autism when using personalised speech generated devices


Background: This study investigates the impact of impairments in declarative memory on the ways children with low-functioning autism (LFA) interact with augmentative and alternative communication (AAC) devices. Recent technological advances in AAC involve attempting to transplant either the natural voice of a minimally verbal child, or a very similar one, onto the device so that the devices’ output will sound more like that of the child (The Nancy Lurie Marks Foundation, 2009). The rationale for developing these features on AAC devices is twofold. Firstly, it is assumed that the user will prefer a more naturalistic output resembling human speech to digitised default speech-output. Second, that both the quantity and quality of the individual user’s communication via the AAC device will be enhanced through a process of reinforcement on foot of conversational partners’ preference for the more naturalistic voice.

Objectives: This first part of this study explores critical components related to voice recognition, levels of familiarity and recollection in a sample of children with LFA compared to age and ability matched control groups. The second part of this study investigates childrens use of personalised speech output options on AAC devices over a three month period.

Methods: Three groups of children were recruited; thirty-three children with LFA, thirty-three typically developing (TD) children matched with the LFA group for non-verbal mental age (NVMA), and twenty-seven children with developmental delay (DD) without autism, matched with the LFA group for chronological age (CA), verbal, and nonverbal ability. An experimental design was used. A series of five tests were conducted to assess voice recognition, levels of familiarity and degrees of recollection across the groups. Based on this data, 12 children with LFA were allocated AAC devices. Language activity monitoring (LAM) software tracked their individual use of the device over a three month period. Analyses undertaken included, tests of association (correlation), tests of difference (t-tests, ANOVAs), and calculations of effect size.

Results: The analyses in part one showed that children with LFA are significantly impaired in voice processing abilities and in relation to familiarity and recollection relative to DD and TD control groups. This trend remained after controlling for gender, NVMA, schools, and CA. The analyses in part two indicated that personalised speech outputs did not serve to increase or enhance communication in this cohort. Rather than preferential use of the personalised speech output, differential use of speech output was evident with partial and complete abandonment of the devices by a small number of the children.

Conclusions: The combined language and learning deficits in LFA make them ideal candidates for AAC systems however, due to perceptual processing and sociocognitive impairments, additional features such as pre-recorded natural voices on AAC devices may be of no added benefit to this group. Instead a series of short, inexpensive tests can be conducted to match the right system to each individual child with a focus on the child’s abilities, preferences, and evolving needs.

499 words

| More